My First FAccT Conference #FAccT2022

TLDR: Attending the ACM FAccT conference in Seoul, South Korea for the first time this year, gave me a bit of hope that there are more of me. By that I mean, more people who are sitting at their computers trying to find answers to impossible questions about how to overturn longstanding systems of power and control, reach a place of fairness, accountability, and transparency, and transfer that to artificial intelligence systems. Much of the work is closely examining the depths of the problem, and it is rigidly systemic. It is beautiful to see the work being done, and so many people in one place who have dedicated their lives to the task.

My First Day at FAcct

On the first day, I was early, but not that early. I set out from my hotel room in the exact opposite direction that I needed to for the conference hall, flagged down a taxi, and hoped I was headed the right way. I couldn’t miss the huge Coex building, which stood as modern as can be, across from a mountain dotted with Korean Buddhist temples and a giant statue. I gestured to the driver to stop, almost speaking to him in Khmer, but realizing I was not in Cambodia, where I have been living and working remotely. I was in Korea!

The statue and the Coex building, from Korea tour guide gives something other than unwinding — Brick Ridge

Being a Volunteer Coordinator

I love to volunteer at events. You get more of the back-stage feel and get to be a part of the inner workings of it all. After getting my paper submission rejected for the conference, I was delighted to still be accepted as a volunteer coordinator. I hoped that it would put me in a position to meet some of the other volunteers and organizers, in addition to helping the event run smoothly overall.

I knew that my co-volunteer coordinator was at the reception desk and that she was looking for me. I told her I had short light-colored hair. (I am never quite sure what my hair color is, it had been light pink but coming out of the shower, it could really be closer to blond now, depending on the light.)

She spotted me right away, a young and bright Korean girl with hair lighter than mine, half tied in a black bow. She smiled at me, we exchanged names and pleasantries, and I told her whatever she needed me to do, I was ready.

She took me through the crowds of people checking in, and down the corridor to the volunteer room, which was a quiet escape from the hustle and bustle of the conference. There we sat down and started sending messages to the volunteers who were on the schedule for the day, a task we would do every day of the conference. In the background, the first keynote presenter was on someone’s screen, talking from just the other room. As a hybrid conference, many of the talks could be attended either in person or virtually, or both.

We fell into a great rhythm together, and on a break, we walked down to the mall on a floor below the conference hall. We felt like kids sneaking off to go shopping. She took me into a photo booth, which is quite popular in Korea, and we got some polaroid pictures taken together. It was great fun.

Our photo-booth pictures at FAacct#2022

The People I Met at FAccT2022

The conference was attended by a mashup of folks from disciplines ranging from computer science to law to philosophy to activism. I didn’t meet any other anthropologists, however, I did meet a couple of people who said they were wannabe anthropologists. What they all had in common was a connection to ethics in artificial intelligence, primarily, as the namesake of the conference says, Fairness, Accountability, and Transparency in sociotechnical systems.

Some of the best conversations happened on the bus ride to get COVID tested, as was required for those just arriving in Korea. They took us in small groups in a van to a testing site. On the way, I met a few interesting people, including a Ph.D. student researching the philosophy of ethics in AI at Georgetown and another from Australia studying the legal angle for ethical AI development. One man from Michigan started asking about my research and expressed that he was there with the purpose to try and find solutions to the many outstanding questions of this industry.

In retrospect, I find this an odd goal in an ethical AI space. Seeking solutions is often what gets us into trouble. A lot of terrible things are done in the name of solutions to big problems. Later on, in a conversation with another guy who commented on my work in the decolonial space. He said that he didn’t really understand decolonization, and asked, was there some end goal to focusing on this topic? Were we all meant to go back to a sort of primitive time before colonization began?

I now notice a theme in these lines of questioning, which tend to come from educated white men, that is inherently problematic for me. It has to do with needing to know the goal or the solution. I am not saying that there are no solutions to the complexity involved in attempting to act from an ethical place in sociotechnical fields: there just isn’t one solution or one goal. This whole solutions-oriented concept to technical ethics applications needs a re-think.

These solution-minded people need to take a seat for a moment and focus on listening, and keep listening. To specifically answer the decolonization question, I would offer that it is more about making corrections and learning from the past at every turn, and not repeating the mistakes of the past, but finding ways to interrupt the systems of oppression that like to categorize and therefore minoritize people, causing further disadvantages to those already struggling in a world that is made to benefit a select few.

This is extremely challenging. Systems of oppression are deeply ingrained into technology and how it works, and as systems become more autonomous, they are learning to oppress people all on their own.

Systems of oppression are deeply ingrained into technology and how it works, and as systems become more autonomous, they are learning to oppress people all on their own.

Thoughts on the Many Sessions at FAccT

I found comfort in my own familiarity with the material being presented. Many of the talks I attended touched on areas I have researched, and many used examples I am familiar with, such as Dr. Joy’s work with facial recognition and the recidivism algorithm-gone-wrong. The concepts of fairness, accountability, and transparency were certainly thematic, particularly in the CRAFT poster sessions. There were a lot of fresh concepts as well, including things that went a bit over my head and littered with computer science jargon.

One of the first panels I attended was about intersectionality, which I was excited about because intersectional feminism is something I have always been interested in. In computing and AI, binaries and categories are common and useful. However, people don’t fit so neatly into these distinctions. And it can be harmful to force them into these boxes for the benefit of data mining and the like.

A really good point that was made on the panel was that it is impossible to always include everyone. No matter what you do, some people or groups will be excluded. It has to be decided to what extent this is tolerated on a case-by-case basis. Another good point was that when it comes to intersectionality, issues can have multiplicative effects and are not just additive. So if a person is female, African American, and disabled, that is a huge multiplication of features that stray from the majoritized group of able-bodied white males.

I use the word majoritized and also minoritized, because white males are not actually the majority. One might think so by looking at all of the research that historically supports them and ignores and invisibilizes all others. So-called minorities are not always the minority, especially globally. Research tends to be Western and generally US-focused, not allowing for the reality that there are whole continents of non-white people who get put into a minority category when the technologies being researched are deployed globally. It makes me a bit mad, but it was great to be in a space where people recognize the same issues.

Final Takeaways

Overall FAcct this year was a space full of shy, geeky, sweet, loving individuals who care about the world and what happens to it, which I just adored! It was sometimes difficult to strike up a conversation with someone, but shells were thin and once eyes met, smiles could be felt under our masks.

In a world that is far from fair, accountable, or transparent, these individuals stand for bringing these qualities into the light, and that is a beautiful thing. Hopefully, with all these bright young minds focused on the complex issues that stem from AI and tech but that radiate out into the world, this conference and its attendees can serve to create the positive impact we’re intending to.

Sometimes it seems like we are up against a giant in tech, and there are a lot of forces trying to keep things as they are, and even go backward. Progress does not come naturally. Just because technology advances, it does not mean it is up with the times. That takes a lot of work. I hope that it can radiate out of the bubbles we make and the research papers we write.

And I can’t wait for another trip to Korea!

With hope and love, Jennafer Shae Roberts

Exploring the streets of Seoul, South Korea

You can stay up to date with Accel.AI; workshops, research, and social impact initiatives through our website, mailing list, meetup group, Twitter, and Facebook.